Cloud deployment models

Common cloud deployment models

Public Clouds

Public clouds are a type of cloud computing that provides storage, compute and networking services, software, and facilities on demand via the internet. Public clouds are multi-tenant environments, meaning multiple customers will share the underlying resources that are owned and operated by the cloud service provider (CSP). Popular public cloud vendors include Amazon Web Services (AWS), Microsoft’s Azure, Google Cloud Platform (GCP), and Oracle Cloud Infrastructure (OCI).

Private Clouds

A private cloud is typified by resources dedicated to a single customer; no other customers will share the underlying resources (hardware and perhaps software). Therefore, private clouds are not multi-tenant environments. Instead, an organization might own and operate a private cloud as the sole customer; or the organization might contract with a cloud provider for exclusive use of specific resources inside what otherwise would be a public cloud; or a cloud provider installs hardware/software inside the customer’s data center – the customer controls the data plane and consumes the resources while the cloud provider owns the hardware/software and takes the ownership of the control plane.

Hybrid Clouds

A hybrid cloud contains elements of the private cloud services, public cloud services, and on-premises infrastructure. For instance, an organization might want to retain some private cloud resources (say, their legacy production environment, which is accessed remotely by their users) but also lease some public cloud space (maybe a Platform-as-a-Service (PaaS) function for software development/testing, away from the production environment to lessen the risk of crashing operational systems).

Multi-Cloud

Multi-cloud is a strategy that leverages two or more cloud services and mainly refers to multiple cloud services. Typically, each cloud is managed by a CSP. Also, some researchers treat hybrid clouds and multi-clouds as the same concept since both cloud deployment models consist of more than a single infrastructure management mechanism.

Community Clouds

A community cloud features infrastructure and processing owned and operated by (or for) an affinity group; disparate pieces might be owned or controlled by individuals or distinct organizations, but they come together in some fashion to perform joint tasks and functions. It can be provisioned by a third party on behalf of the various members of the community, such as government customers that have similar security/service requirements while needing to be isolated from common public clouds. The gaming industry could be another group of customers who are in favor of community clouds, such as Sony’s PlayStation network that involves many different entities coming together to engage in online gaming.

Cloud deployment model comparison

When enterprises prepare to migrate to clouds, it will be beneficial to review the different cloud deployment models based upon business needs, risks associated with each type, and best practices for certain industries.

Choosing a cloud deployment model

Choosing a cloud deployment model comes down to a series of trade-offs related to cost, management, and security. While public clouds may be the best option for small organizations from a cost perspective, organizations that require more control and/or security may opt for a private or hybrid cloud, provided they have the manpower and budget to manage those deployments effectively. Furthermore, it’s critical to consider your application architecture as well.

Undoubtedly, no single deployment model fulfills all of an organization’s needs. Choosing the most suitable cloud deployment model for an organization will depend upon the business requirements. Here are some criteria to consider:

IT costs

Shifting to cloud computing may considerably decrease the cost of managing and maintaining your IT systems since organizations need not buy and install expensive systems and equipment and convert CAPEX to OPEX. As previously illustrated, public cloud and multi-cloud will be the most feasible deployment models in this regard. For example, to build a disaster recovery (DR) site on a private cloud would cost more compared to public or multi-cloud since the organization would have to pay double; whereas putting a DR site on a public cloud would require very little infrastructure and would more easily spin up resources if disaster happens.

Risk tolerance

Organizations need to evaluate the risk tolerance and choose the right models for the business. For example, government agencies will likely place the mission-critical workloads in government clouds instead of commercial clouds, even if both belong to the same CSP. One of the main reasons is that the dedicated government cloud would have higher security standards than commercial one.

Scalability

Organizations in industries such as retail or life sciences need high computing power to process data in a short period of time. It is not cost effective and time consuming to do so on private/hybrid or even community cloud.

Business continuity

Organizations often need to consider the impact of legislative environment changes. Such changes could greatly impact the business. Having an environment with readiness to spin up to full scale will be critical to business continuity if legislative environments change. Obviously public cloud and multi-cloud models will be the best in terms of time and costs.

Based upon the criteria listed above, multi-cloud is undoubtedly the winner as it has the most advantages that organizations need for their day-to-day business operations. In fact, many reports indicate that most organizations are adopting a multi-cloud strategy, and more organizations will follow suit moving into the future. Providers will also seek to create partnerships that combine their mutual strengths to speed up market launches and time to market for multi-cloud products and services.

However, multi-cloud also has some unique challenges as it is more complex compared to using a single CSP for deployment as well as day-2 operations. For organizations who would like to standardize the design across different CSPs, adopt orchestration mechanisms to manage cloud environments, enhance visibility, and secure cloud networking to the next level, Aviatrix will be able to help.